On Wed, Mar 05, 2025 at 07:10:05PM -0800, Eric Biggers wrote: > > Why? All the callers keep track of the address anyway. I don't see a need to > bloat the scatter_walk structure beyond a simple (sg, offset) pair. Because the existing convention is error-prone and relies on the caller keeping the mapped pointer intact. But I will consider your objection about bloating the struct. The places that I've changed are simply storing the object on the stack anyway, so there is no actual bloat. Do you have a usage scenario where this struct is embedded into something bigger and bloat is a real concern? Thanks, -- Email: Herbert Xu <herbert@xxxxxxxxxxxxxxxxxxx> Home Page: http://gondor.apana.org.au/~herbert/ PGP Key: http://gondor.apana.org.au/~herbert/pubkey.txt
